MoDist: Motion Distillation for Self-supervised Video Representation Learning

Published in arXiv preprint, 2021